antsnormflows.nets.lipschitz

Functions

asym_squash(x)

leaky_elu(x[, a])

normalize_u(u, codomain[, out])

normalize_v(v, domain[, out])

projmax_(v)

Inplace argmax on absolute value.

vector_norm(x, p)

Classes

InducedNormConv2d(in_channels, out_channels, ...)

InducedNormLinear(in_features, out_features)

LipschitzCNN(channels, kernel_size[, ...])

Convolutional neural network which is Lipschitz continuous with Lipschitz constant L < 1

LipschitzMLP(channels[, lipschitz_const, ...])

Fully connected neural net which is Lipschitz continuou with Lipschitz constant L < 1

Swish()